extend the dict-TTS concept to newer TTS models like metavoice TTS and Tortoise TTS

[StephennFernandes]

hi given that there are newer TTS models like metavoice that 1B parameters and have exceptional high quality speech. along with Tortoise TTS whats the scope that this implementation can be extended to newer TTS models.

FYI: i am asking for the technical possibility and some rough guidance so i could take this forward and/or submit a PR.